Teen Mothers’ Educational Attainment and their Children’s Risk for Teenage Childbearing

C Emily Hendrick; Julie Maslowsky

doi:10.1037/dev0000705

. Author manuscript; available in PMC: 2020 Jun 1.

Published in final edited form as: Dev Psychol. 2019 Feb 28;55(6):1259–1273. doi: 10.1037/dev0000705

Teen Mothers’ Educational Attainment and their Children’s Risk for Teenage Childbearing

C Emily Hendrick ¹, Julie Maslowsky ²

PMCID: PMC6533138 NIHMSID: NIHMS1012673 PMID: 30816724

Abstract

The children of teen mothers are at elevated risk for becoming teen parents themselves. The current study aimed to identify how levels of mothers’ education were associated with risk of teenage childbearing for children of teen versus non-teen mothers. Through structural equation modeling, we tested whether children’s environmental and personal characteristics in adolescence and subsequent sexual risk behaviors mediated the relationship between their mothers’ educational attainment and their risk for teenage childbearing. With multiple-group models, we assessed whether the associations of maternal educational attainment with children’s outcomes were similar for the children of teen and non-teen mothers. The sample (N = 1,817) contained linked data from female National Longitudinal Survey of Youth 1979 (NLSY79) participants and their first-born child (son or daughter) from the NLSY79 Children and Young Adults. The mediating pathways linking higher levels of maternal education to lower risk for teenage childbearing, and magnitudes of the associations, were mostly similar for children of teen and non-teen mothers. However, non-teen mothers experienced greater associations of their high school diploma attainment (versus no degree) with some of their children’s outcomes. Also, the association of earning a high school diploma (versus a GED) with household incomes was greater for non-teen mothers; there was no significant difference between degree types for teen mothers. Findings provide support for teen mother secondary school support programming, but point to a need for further research regarding the long-term behavioral and social outcomes associated with the high school equivalency certificate for teen mothers and their children.

Keywords: maternal education, teenage childbearing, intergenerational cycle of teenage childbearing

The sons and daughters of teen mothers are at elevated risk for becoming teen parents themselves (Campa & Eckenrode, 2006; East, Reyes, & Horn, 2007; Jutte et al., 2010; Kim, 2014; Meade, Kershaw, & Ickovics, 2008; Pogarsky, Thornberry, & Lizotte, 2006). This phenomenon, known as the intergenerational cycle of teenage childbearing (Meade et al., 2008), is concerning due to the cycle of health and socioeconomic adversity associated with teenage childbearing for teen parents (Assini-Meytin & Green, 2015; Henretta, 2007) and their families (Jutte et al., 2010).

Children of teen mothers are at elevated risk for teenage childbearing due to a constellation of factors: they more often grow up in lower-resource settings, with lower quality parent-child interactions, and have lower academic aptitude compared to children born to older mothers (Gibb, Fergusson, Horwood, & Boden, 2015; Lewin, Mitchell, & Ronzio, 2013; Tang, Davis-Kean, Chen, & Sexton, 2014). In turn, relatively lower-resource settings, lower-quality parent-child interactions, and poor academic aptitude each place children at increased risk for sexual behaviors in adolescence that lead to teenage pregnancy and childbearing (Carlson, McNulty, Bellair, & Watts, 2014; Huebner & Howell, 2003; Kirby & Lepore, 2007).

Maternal education is a strong predictor of healthy development and well-being for children (Perper, Peterson, & Manlove, 2010). Teen mothers have historically attained low levels of education (Kane, Philip Morgan, Harris, & Guilkey, 2013). Accordingly, it is possible that increases in teen mothers’ educational attainment could help to create developmental contexts for their children that reduce their children’s risk for teenage childbearing. In the current study, we assess the mechanisms by which mothers’ education is associated with their children’s risk for teenage childbearing, and whether these mechanisms, and their magnitudes, vary by mothers’ age at childbearing.

The Intergenerational Cycle of Teenage Childbearing

Teenage childbearing in the United States has declined over 50% since 2007 (Martin, Hamilton, Osterman, Driscoll, & Drake, 2018). The recent decline has been attributed to increases in effective contraception use among adolescents (Lindberg, Santelli, & Desai, 2016). However, vast disparities in teen birth rates remain across geographically, economically, and racially-defined subpopulations within the United States (Martin et al., 2018). The offspring of teen parents are a subpopulation at especially elevated risk for early childbearing. The children of teen mothers have demonstrated rates of early childbearing approximately 2–3 times those of children born to older mothers (Jaffee, Caspi, Moffitt, Belsky, & Silva, 2001; Meade et al., 2008). Although the daughters of teen mothers are most often considered in research of the intergenerational cycle of teenage childbearing, sons of teen mothers area at increased risk for teenage childbearing as well (Barber, 2001; Jaffee et al., 2001; Kim, 2014; Pogarsky et al., 2006; Wildsmith, Manlove, Jekielek, Moore, & Mincieli, 2012). Also, a study of national cohorts of women born from the 1950’s through the early 1980’s, who first gave birth in the 1960’s through the 2000’s, observed that women who were offspring of teen mothers experienced their first births at increasingly younger ages over time, while the children born to women over the age of 25 experienced childbearing at increasingly older ages (Kim, 2014). This increasing disparity points to a specific and growing need to identify salient strategies for reducing the risk of teenage childbearing among the children of teen mothers.

Teen Mothers’ Education as a Potential Pathway for Reducing the Intergenerational Cycle of Teenage Childbearing

According to social ecological models of the influences of adolescent health behaviors and health (DiClemente, Salazar, Crosby, & Rosenthal, 2005), adolescents’ sexual behaviors and risk for childbearing are influenced by the norms and resources in their neighborhoods, schools, and homes, as well as their interpersonal interactions with adults, peers, and family members (Meade et al., 2008). Each of these influences on adolescents’ reproductive and sexual health and behaviors is potentially influenced by their mothers’ educational attainment (Figure 1). Consequently, there are several potential mechanisms through which maternal education can promote positive health behaviors and health of children, including their sexual and reproductive health in adolescence (Figure 2).

Figure 1. — Social ecological influences on children’s risk for teenage childbearing potentially impacted by increases in maternal educational attainment.

Figure 2. — Conceptual model of the mechanisms by which mothers’ (G1) education level predicts children’s (G2) risk for teenage childbearing.

Through Children’s Environmental Contexts

A woman’s educational level is associated with the environmental contexts in which her children develop. Mothers with higher educational attainment generally have higher paying jobs, live in healthier and safer neighborhoods, couple with partners with higher educational attainment, have lower rates of divorce, and are more likely to have access to quality healthcare for themselves and their children (Fry & Taylor, 2012; Greenwood, Guner, Kocharkov, & Santos, 2014; Strickland, Jones, Ghandour, Kogan, & Newacheck, 2011; US Bureau of Labor Statistics, 2014). Thus, the offspring of mothers with higher educational attainment are more likely to grow up in higher-resource households, schools, and neighborhoods that position children on healthy developmental trajectories through childhood and adolescence. Accordingly, youth from higher-resource settings demonstrate fewer sexual risk behaviors in adolescence (Carlson et al., 2014) and have better access to reproductive health care services (Hall, Moreau, & Trussell, 2012).

Few studies have directly assessed how teen mothers’ education may influence their children’s environmental contexts, which, in turn, may lead to reductions in their children’s risk for teenage childbearing. In one example, although not directly assessing teen mothers’ education, Wildsmith and colleagues assessed the associations among teen mothers’ early characteristics (including academic aptitude), their children’s home environments, their children’s risk behaviors in adolescence, and subsequent risk for teenage childbearing using linked data from the mothers and children of the National Longitudinal Survey of Youth 1979 cohort (NLSY79) and the NSLY79 Children and Youth Adults (NSLY79CYA) (2012). They found that early maternal characteristics influence their children’s risk for 2nd generation teenage childbearing through the household context and their children’s risk behaviors in adolescence. However, the authors did not include the children’s sexual risk behaviors in adolescence among the risk behaviors examined in the study—a notable exclusion considering that sexual risk behaviors and other risk behaviors do not always go hand in hand in adolescence (Sullivan, Childs, & O’Connell, 2010).

Through More Involved Parenting

Maternal educational attainment is also linked to the quality of interpersonal interactions between mother and child. Mothers with higher educational attainment exhibit more involved and supportive parenting practices such as interacting with children’s school environments, demonstrating positive discipline practices, and creating a supportive caregiving environment at home (Crosnoe, Ansari, Purtell, & Wu, 2016; Domina & Roksa, 2012; Magnuson, 2007). This higher level of parental involvement is potentially partially attributed to women’s educational experiences. For example, women may learn the importance of positive parenting practices for healthy child development through their educational experiences and exposure to curriculum. Also, women may emulate positive parenting practices of parents they encounter through, and as a result of, their educational experiences. In fact, when mothers attain additional years of formal schooling after their children are born, their parenting practices improve (Domina & Roksa, 2012).

More involved parenting, such as higher levels of parental monitoring, in turn, can protect against some adolescent sexual risk behaviors such as younger initiation of sex, unprotected sex, and having more sexual partners (Huebner & Howell, 2003; Li, Stanton, & Feigelman, 2000; Luster & Small, 1994; Rai et al., 2003). Moreover, in a study using data from the NLSY 1997, Meade and colleagues (2008) identified child-reported low parental monitoring as a salient predictor of teenage childbearing specifically for the children of teen mothers. While perceived parental monitoring levels were not associated with risk for teenage childbearing for the daughters of older mothers, the daughters of teen mothers had a 63% increased risk of having a teen birth if they reported low levels of parental monitoring compared to the daughters of teen mothers who reported high parental monitoring.

Through Children’s Own Academic Aptitude and Performance

Mothers’ educational attainment is strongly positively linked to their children’s academic aptitude and performance (Carneiro, Meghir, & Parey, 2013; Harding, 2015; Magnuson, 2007; Tang et al., 2014). This relationship has been observed among children in general, as well as specifically among the children of teen mothers (Tang et al., 2014). Maternal educational attainment is believed to influence children’s academic outcomes through several mechanisms. For example, Harding and colleagues propose that the gains in human, cultural, and social capital women experience with higher levels of education indirectly impact their children’s proximal experiences, and ultimately, their academic outcomes (Harding, Morris, & Hughes, 2015). Their theoretical framework posits that these relationships are mediated through a variety of maternal mechanisms including language use in the home, homework help, and educational support resource-finding, among others. In turn, youth with higher academic performance are at reduced risk for sexual risk behaviors in adolescence and teenage pregnancy and childbearing (Kirby & Lepore, 2007; Lou & Thomas, 2015). In contrast, youth who struggle academically and drop out of high school begin childbearing earlier than those who persist in high school and beyond (Manlove, 1998).

Teenage Childbearing among Children of Teen Mothers versus Children of Non-Teen Mothers

To our knowledge, only one study to date has examined the predictors of teenage childbearing for the children of teen versus older mothers (Meade et al., 2008). Meade and colleagues (2008) found lower parental monitoring, deviant peer norms, and household poverty to predict teenage childbearing only for the daughters of teen mothers. Conversely, they found the timing of menarche to predict teenage childbearing only for the daughters of older mothers. As this is the only known study to assess whether the predictors of teenage childbearing differ for the offspring of teen and non-teen mothers, additional research is needed to elucidate whether the mechanisms by which the risk for teenage childbearing is elevated or reduced varies for the children of teen and older mothers.

The Resource Substitution Theory

The resource substitution theory posits that individuals from lower-resource settings experience greater educational returns on their later-life health and well-being because education acts as a “substitute” for other health-promoting resources they may lack compared to individuals from higher-resource settings (Ross & Mirowsky, 2006). Consequently, individuals from higher-resource settings do not experience as large educational returns because their health and well-being are already supported by their pre-existing resources, regardless of the amount of education they attain. This theory has been applied to explain differential educational returns on health and well-being by gender and socioeconomic status (Bauldry, 2015; Ross & Mirowsky, 2011) and offers one explanation for why we may expect to see differing magnitudes of associations of educational attainment with later-life outcomes for teen mothers as compared to non-teen mothers: teen mothers often come from lower-resource settings than non-teen mothers. Accordingly, younger mothers from more disadvantaged backgrounds have shown greater gains in positive parenting practices and in their children’s educational outcomes from increases in educational attainment when compared to older mothers from more advantaged background (Magnuson, 2007). Thus, one would expect teen mothers’ education to have larger correlations with their household incomes, parenting practices, and their children’s academic and behavioral outcomes than non-teen mothers. However, to our knowledge, the resource substitution theory has not yet been tested among a sample of teen versus non-teen mothers.

Teen Mothers Earn GEDs at Higher Rates

Teen mothers earn GEDs at higher rates than women who delay childbearing beyond the teenage years (Perper et al., 2010). While, on average, the later-life health and socioeconomic outcomes of GED earners more resemble those with no degree than those who earn a high school diploma (Cameron & Heckman, 1993; Zajacova & Everett, 2014), it is not known whether the outcomes associated with the GED for teen mothers are equivalent to those of the general population. Some pregnant and parenting teens report being steered toward alternative secondary schooling programs (Pillow, 2004) that may result in a GED in lieu of a high school diploma. As some students who would have otherwise attained high school diplomas instead earn GEDs due to their unique challenges as pregnant or parenting students, earning a GED due to pregnancy or parenting versus other reasons may change the later significance of the GED. In addition, the academic experience (e.g., curricula, peers, and teachers) in educational programs for pregnant and parenting students may differ from those of programs for students who are not pregnant nor parenting. For example, programs for pregnant and parenting students often include curricula for delaying additional pregnancies and developing parenting and other life skills (Costello & Institute for Women’s Policy Research, 2014). Thus, the qualitative experience of earning a GED may differ for teen and non-teen mothers. Consequently, the GED may have differing associations with teen mothers’ and their children’s later life outcomes than with those of women who delay childbearing until after the teenage years.

The Current Study

In the present study, we prospectively examine the associations of maternal education with offspring’s childbearing in a longitudinal, multigenerational sample. We build upon Wildsmith and colleagues’ research (2012) in this area by specifically assessing the association of maternal educational attainment with mechanisms for reducing children’s risk for teenage childbearing. We also examine whether these mechanisms vary for the children of teen versus non-teen mothers.

Further, the current study assesses teen mothers’ degree type (no degree v. GED v. high school diploma) as the maternal educational attainment predictor of interest. As teen mothers attain GEDs at higher rates than women who delay childbearing, and the health and social benefits of the GED have been demonstrated to be less than those of the high school diploma, it is important to distinguish this degree type when assessing the relationships between teen mothers’ education and her and her children’s long-term outcomes, including teenage childbearing. We include both the sons and daughters of teen mothers, advancing previous literature that primarily examines predictors of daughters’ risk for teenage childbearing even though the sons of teen mothers are also at elevated risk for teenage childbearing (Kim, 2014). Finally, we control for a wide range of maternal background characteristics often not accounted for in studies of maternal education and subsequent child outcomes. These background factors represent the personal, environmental, and academic experiences of mothers in adolescence that could potentially influence their educational outcomes and, indirectly, their children’s risk behaviors in adolescence that could lead to teenage childbearing.

The current study aims to address the following research questions:

Are the mechanisms (household income, perceived parental monitoring, and/or academic aptitude, and subsequent sexual risk behaviors in adolescence) by which higher levels of maternal education are associated with lower risk of teenage childbearing different for the children of teen and older mothers? In other words, does teen motherhood status moderate the mechanisms by which mothers’ education is linked to children’s predicted probability of teenage childbearing? Drawing from social ecological models, we posit that each of the tested mechanisms (household income, perceived parental monitoring, and academic aptitude, and subsequent sexual risk behaviors in adolescence) will partially mediate the relationship between both teen and non-teen mothers’ educational attainment and their children’s risk for teenage childbearing.
Are the mechanisms driving this association of the same magnitude for the children of teen and older mothers? In other words, does teen motherhood status moderate the magnitude of the mechanisms by which mothers’ education is linked to children’s predicted probability of teenage childbearing? Drawing from the resource substitution theory, we hypothesize that the associations of maternal educational attainment with the tested mechanisms (household income, perceived parental monitoring, academic aptitude, and subsequent sexual risk behaviors in adolescence) will be greater for teen mothers than non-teen mothers.

Methods

Data Sources

Data for the current study were from linked mother-child pairs from the NLSY79 and NLSY79CYA from 1979 through 2014. As data were de-identified and publicly available, the current study was deemed not to be human subjects research by the institutional review board at The University of Texas at Austin.

National Longitudinal Survey of Youth 1979 (NLSY79).

NLSY79 is an ongoing longitudinal project of the US Bureau of Labor Statistics that surveyed US youth ages 14–22 in 1979 and annually through 1994, then biennially thereafter (US Bureau of Labor Statistics, n.d.-b). Topics covered in the NLSY79 include education, training, employment, household, family background, romantic partnerships, fertility, children, income, public program participation, health, attitudes, expectations, non-cognitive tests, activities, crime, and substance use.

National Longitudinal Survey of Youth 1979 Children and Young Adults (NLSY79CYA).

NLSY79CYA began in 1986 and is a separate biennial survey of all children born to NLSY79 female participants. Children under age fifteen complete a “child” version of the survey that includes information from the children and mothers regarding the children’s environment, development, behaviors, health, and interpersonal relationships. Starting the year child participants turned 15 years old, they begin completing a more comprehensive “young adult” survey that paralleled the NLSY79 survey in many ways, such as the inclusion of sexual and reproductive health items (US Bureau of Labor Statistics, n.d.-c).

The analytic sample.

The NLSY79 participants are referred to as Generation 1 (G1) and NLSY79CYA participants are referred to as Generation 2 (G2), for clarity. Linked NLSY79 and NLSY79CYA datasets were a fitting source of data for the current study as together they provided rich information about children’s (G2) experiences through adolescence as well as information about mothers’ (G1) background characteristics that may directly or indirectly influence mothers’ (G1) educational experiences, parenting behaviors, children’s household environments (G2), and children’s health behaviors and outcomes (G2). Figure 3 illustrates the process by which we arrived at the current analytic sample (N = 1,817 mother (G1)-child (G2) pairs). We began with the full sample of 11,521 children (G2) of NLSY mothers (G1) and then excluded those who were not first-born children (G2) (n = 6,591). We then excluded pairs in which the child (G2) was born prior to the mother’s (G1) Wave 1 interview (n = 986) due to timing of reporting of potentially confounding maternal (G1) background characteristics measured at Wave 1 occurring before and after the birth. We then excluded those pairs in which the mother (G1) was age 19 or older at the Wave 1 interview to preserve temporality between baseline characteristics and mothers’ (G1) educational outcomes. Finally, we excluded pairs in which children (G2) were younger than 19 in their most recent interview (n = 765) as they may have had a teen birth after their most recent interview. The NLSY79 does not recommend using sampling weights when conducting regression analyses for subpopulations (US Bureau of Labor Statistics, n.d.-a). Therefore, the current study does not use sampling weights in analyses and does not assume a nationally representative sample, but rather includes a racially and ethnically diverse sample of mothers (G1) and their first-born children (G2) in the United States.

Figure 3. — Creation of analytic sample including 1,817 NLSY79/NLSY79CYA mother/child pairs.

Measures

Unweighted descriptive characteristics of the analytic sample are displayed in Table 1. G2 consisted of 49–50% sons and 50–51% daughters for both G1 teen and non-teen mothers.

Table 1.

Unweighted Descriptive Characteristics of Analytic Sample

	Generation 1 all mothers N = 1,817		Generation 1 teen mothers n = 323		Generation 1non-teen mothers n = 1,494

	Mean or Proportion	SD	Mean or Proportion	SD	Mean or Proportion	SD
G2 characteristics

Gender
Girls	0.51		0.50		0.51
Boys	0.49		0.50		0.49
First birth prior to age 19	0.06		0.14		0.04		^***
Age at first birth
All	21.89	3.88	21.71	4.26	21.97	3.70
Girls	21.38	3.98	21.29	4.42	21.42	0.80
Boys	22.59	3.62	22.22	4.02	22.77	3.40
Age at first sex
All	16.15	2.50	14.87	2.55	16.43	2.40	^***
Girls	16.45	2.28	15.62	2.03	16.63	2.29	^***
Boys	15.84	2.67	14.14	2.79	16.23	2.49	^***
Reported unprotected sex during adolescence	0.21		0.32		0.19		^***
Household income at age 14 (divided by $10,000)	5.11	5.81	2.63	2.27	5.65	6.19
Perception of parental monitoring at age 14
Mom sometimes or hardly ever knows who youth is with	0.19		0.19		0.19
Mom often knows who youth is with	0.81		0.81		0.81
PIAT math percentile	52.20	27.97	40.10	25.01	54.83	27.89	^***
PIAT reading comprehension percentile	46.43	26.87	36.33	25.10	48.64	26.75	^***
PIAT reading recognition percentile	59.17	29.16	48.48	29.59	61.50	28.55	^***

G1 characteristics

First birth prior to age 19	0.18
Degree type attained by age 30
No degree	0.08		0.20		0.06		^***
GED	0.12		0.26		0.09		^***
HS diploma	0.78		0.51		0.84		^***
Years of school completed by age 30	12.79	2.17	11.42	1.91	13.08	2.10	^***
G1 participants’ mothers’ highest grade completed	10.53	3.15	9.44	2.99	10.76	3.14	^***
Enrolled in school at Wave 1	0.88		0.84		0.89		^*
Aptitude percentile score at Wave 1	39.23	27.42	25.18	20.72	42.25	27.75	^***
Highest grade expected to earn at Wave 1	13.73	2.19	12.62	2.03	13.98	2.15	^***
Race/ethnicity
Black/Non-Hispanic	0.31		0.45		0.28		^***
Hispanic	0.20		0.25		0.19		^*
Non-Black/Non-Hispanic	0.49		0.30		0.53		^***
Self-esteem score 1980	21.68	3.98	20.67	3.76	21.89	3.99	^***
Number of siblings at Wave 1	3.90	2.67	4.51	2.90	3.77	2.60	^***

Open in a new tab

p < .05

^**

p < .01

^***

p < .001.

Note. GED: General Education Development/High School Equivalency Certificate. PIAT: Peabody Individual Achievement Test. G1: Generation #1. G2: Generation #2. Two-tailed t-tests conducted for continuous characteristics and χ2 tests conducted for categorical characteristics to determine statistical significance of differences in means and proportions of characteristics by G1 teen motherhood status. Race/ethnicity of G2 participants is reported in the NLSY79CYA as the race/ethnicity of the mothers (G1) from the NLSY79.

G2 teenage childbearing.

The dependent variable of interest was whether children (G2) first became parents prior to age 19. While the children of teen mothers and non-teen mothers had similar overall average ages at first birth, 14% of children of teen mothers experienced a teen birth compared to only 4% of non-teen mothers.

G1 educational attainment.

Participants in the NLSY79 were asked, “Do you have a high school diploma or have you ever passed a high school equivalency or GED test?” at each wave. Those answering affirmatively were then asked, “Which do you have, a high school diploma or a GED?” Those reporting not having received either degree by age 30 were defined as having “no degree” in the current study. In the analytic sample, 51% of teen mothers and 84% of non-teen mothers earned a high school diploma by age 30. Accordingly, 26% and 20% of teen mothers earned a GED and no degree, respectively, compared to 9% and 6% of non-teen mothers (all differences significant at p < .001). As not all women reported educational outcomes when they were exactly age 30 due to interview scheduling and missed interviews, educational outcomes reported at the age closest to age 30 (up to age 30, but not younger than 25) were used. The mean age of G1 participants when reporting educational outcomes was 29.9 regardless of teen motherhood status (teen mothers: M = 29.84, SD = .38; non-teen mothers: M = 29.88, SD = .38). Still, the age at which women reported their educational outcomes (ages 25–30) was controlled for in all models. Some G1 participants went on to complete 16 or more years of formal schooling (4 years of college or higher) after attaining a GED or a high school diploma (n = 5 (2%) of G1 teen mothers and n = 275 (18%) of G1 non-teen mothers). Due to the specific interest in assessing the GED as a predictor of long-term outcomes for teen mothers and their children, and the fact that so few teen mothers earn a college degree, mothers’ (G1) educational attainment was defined as having earned a high school diploma, a GED, or no degree by age 30. Participants who completed additional years of schooling beyond their high school diploma or GED were categorized according to the highest high school equivalency (high school diploma or GED) degree they report earning by age 30. A sensitivity analysis was conducted excluding the mother-child pairs of which the mother completed 4 or more years of schooling beyond high school. Results were substantively similar to those including the complete analytic sample. Therefore, results are presented from the complete analytic sample.

Mediators: G2 environmental and personal characteristics at age 14.

Three variables measured when children were approximately 14 years of age were assessed as potential mediators of the relationship between mothers’ educational attainment and their children’s sexual risk behaviors in adolescence and subsequent risk for teenage childbearing: mothers’ household income, children’s perceptions of parental monitoring, and children’s academic aptitude (Figure 2). As the NLSY79CYA is administered biennially, it was not possible to capture all G2 environmental and personal characteristics at age 14. For each participant, values from their interview closest to age 14 (but no more than 2 years prior to or later) were recorded as environmental and personal characteristics occurring at approximately age 14.

Household income.

Household income was represented with the mothers’ (G1) report of the total family income when youth were approximately age 14. This variable was a summary measure of income from all household members related to the mother by blood or marriage including employment, child support payments, government assistance, and other sources. Household income was log transformed for analyses to address skewness and to accommodate variance maximums allowed in MPlus.

Perceptions of parental monitoring.

Children’s perceived parental monitoring was assessed via the children’s rating of the frequency with which they believed their mothers know who they were with when they were not home: “often,” “sometimes,” or “hardly ever.” The variable was dichotomized into high (1) = “often” (81%) and low (0) = “sometimes or hardly ever” (19%).

Academic aptitude.

A latent variable was created to represent G2 academic aptitude at approximately age 14 using normalized, age-specific percentile scores from three subtests of the Peabody Individual Achievement Test (PIAT): math, reading recognition, and reading comprehension (Cronbach’s α = .82, M = 46.43 – 59.17, SD = 26.87 – 29.16 for all three tests). The PIAT is a commonly-used assessment of children’s academic ability with high reliability and validity (Markwardt, 1989).

Mediators: G2 sexual risk behaviors in adolescence.

Two variables were assessed as potential mediators of the relationships between G2 environmental and personal characteristics at age 14 and G2 risk for teenage childbearing: age at first sex and reporting unprotected sex prior to age 19.

Age at first sex.

In the child and youth surveys, participants ages 13 and older reporting ever having “had sexual intercourse” report their age at which they “first had sexual intercourse.” For the current study, age at first sex was re-coded such that 1 = “age 14 or younger” 6 = “age 19 or older” and values 2–5 represent ages 15 through 18 (M = 3.21, SD = 1.67). This variable was recoded to address outliers in the data, to preserve the temporality in the structural equation model, and to account for the fact that individuals reporting sexual debut at age 19 or later were at equal risk for reporting a teen birth in the present study (i.e., no risk).

Unprotected sex.

Youth who reported having sexual intercourse in the last month or living with a spouse or partner were asked, “The most recent time you had sexual intercourse, did you and your partner use any birth control methods?” Youth answering “no” to this question prior to age 19 were coded as reporting unprotected sex = 1 (yes). Those reporting no sexual activity, those reporting sexual activity but not reporting unprotected sex prior to age 19, and those reporting their age at first sex as age 19 or later were all coded as 0 (no unprotected sex). Twenty-one percent (21%) of youth reported unprotected sex prior to age 19. Thirty-two (32%) of those born to teen mothers reported unprotected sex compared to 19% of youth born to non-teen mothers.

Moderator: G1 Teenage Motherhood Status.

An aim of the current study was to assess whether the relationship between teen mothers’ (G1) educational attainment and their children’s (G2) risk for teenage childbearing is significantly different from that of mothers (G1) who begin childbearing after the teenage years. As such, the moderating variable of interest was teen motherhood status. Teenage motherhood in the current study was defined as giving birth to their first child prior to age 19 (n = 323, 18%). While national statistics include births through age 19 in tabulations of teen births, only births through age 18 were considered teen births in the current study in an attempt to identify women as teen mothers who experienced the pregnancy resulting in their first birth during the stage of life when youth are traditionally attending secondary school.

Control variables.

To assess the relationship between teen mothers’ education and their children’s risk for teenage childbearing, potentially confounding factors were included in all models. Control variables included: G1 age at reporting educational outcomes, G1 race/ethnicity (Non-Hispanic/Black (31%), Hispanic (20%), Non-Black/non-Hispanic (49%)), G1 participants’ mothers’ highest grade completed, G1 self-esteem score in 1980, G1 AFQT aptitude test score percentile, G1 educational expectations—how many years of schooling they expected to complete in their lifetime in 1979, whether the G1 participant was enrolled in school in 1979, and G1 participants’ number of siblings in 1979. Due to the size of the teen mother sample (n = 323), and the large number of parameters estimated in single-group mediation models, a principal component analysis (PCA) of a polychoric correlation matrix of all control variables was conducted in Stata 14 to reduce the number of control variables included in analytic models (Coley et al., 2016). PCA of a polychoric correlation matrix was used as the control variables considered included categorical as well as continuous variables (Kolenikov & Angeles, 2009). PCA with an orthogonal varimax rotation of component loadings resulted in 2 principal components with Eigenvalues greater than 1 that together explained 56% of the variance of all control variables. Principal component scores for the 2 retained components were generated for each participant and used in analytic models. As G1 participants were ages 14–19 at baseline, when potentially confounding factors were measured, G1 age at their baseline interview was additionally controlled for in all analytic models. To address concern that results could possibly reflect stability of risk behavior from childhood to adolescence, we conducted a sensitivity analysis including the first-reported, within-sex, standardized total behavior problem index (BPI) score from NLSYCYA for each G2 as a covariate included in the creation of the principal component scores included in all analytical models. The BPI consists of dichotomized answers to a series of questions G1 participants are asked regarding their G2 participants’ problematic behaviors starting when G2 participants are at age 4. Questions reflect the domains of antisocial behavior, anxiousness/depression, headstrongness, hyperactivity, immature dependency, and peer conflict/social withdrawal (US Bureau of Labor Statistics, n.d.-d). As results did not significantly nor substantively change with the inclusion of this variable, results from the original analyses are presented throughout the manuscript.

Analytic Approach

All data were prepared, and preliminary analyses were conducted, in Stata 14 to generate descriptive statistics (Table 1). Single and multiple-group structural equation modeling (SEM) was conducted in MPlus version 8 using the weighted least squares with means and variance adjusted (WLSMV) estimator, a robust estimator appropriate for models including a combination of continuous and binary dependent variables (Muthen & Muthen, 2012). For continuous dependent variables (household income, academic aptitude, age at first sex), unstandardized linear regression coefficients and standard errors are reported; for binary dependent variables (parental monitoring, unprotected sex, and teenage childbearing), unstandardized probit regression coefficients and standard errors are reported. Positive and negative probit coefficients indicate that an increase in the predictor is associated with an increase and decrease in the predicted probability of the outcome, respectively. Unstandardized probit coefficients represent a change in z-score of the predicted probability of experiencing the outcome for every one-unit change in the predictor. In addition to reporting unstandardized coefficients, standard errors, direction (positive or negative), and statistical significance (at α = .05) of path coefficients, we provide practical interpretations of select linear regression coefficients and of the R² for select probit regressions within the final multiple-group structural equation model. An R² for a probit model represents the proportion of the variance explained in the underlying continuous latent outcome variable associated with the observed binary outcome variable.

With the WLSMV estimator, pairwise present analysis is used, and each path coefficient is estimated with all observations with data available for the specific path. With this method, only observations with missing data on exogenous variables or with missing data on all dependent variables are excluded from models, of which there were none in the analytic sample. Missing data ranged from 0% to 16% on variables included in analytic models with items concerning G2 sexual risk behaviors in adolescence missing the most data (15–16%), followed by G2 environmental and personal characteristics at age 14 (6–15%). Model fit was determined for each model using several fit indices: a Root Mean Square Error of Approximation (RMSEA) value of less than .06, Comparative Fit Index (CFI) and Tucker Lewis Index (TLI) values of greater than .95, and a Weighted Root Mean Square Residual (WRMR) value of close to 1.0 all suggest good model fit (Schreiber, Nora, Stage, Barlow, & King, 2006; Yu, 2002). For nested models, chi-square difference tests were used to determine whether the less parsimonious model (i.e., the model with fewer paths constrained to equality across groups in multiple-group models) significantly improved model fit when compared to the more parsimonious model at α=.05.

In preliminary analyses, a confirmatory factor analysis was estimated to test for measurement invariance of the latent variable (academic aptitude) across the two groups. In the model, the factor loadings of each of the three test scores of the latent variable were constrained across groups to ensure “academic aptitude” represents the same construct for both teen mothers and non-teen mothers prior to considering it as a variable of interest in the multiple-group structural models (Keith, 2006). In addition to the fit indices previously described, a Standardized Root Mean Square Residual (SRMR) was provided with the confirmatory factor analysis. An SRMR value of less than .08 is generally considered to represent good model fit (Hu & Bentler, 1999) and all indices suggested that the constrained model fit the data well (RMSEA = .043, CFI = .995, TLI = .995, Standardized Root Mean Square Residual (SRMR) = .025). The model with all factor loadings constrained did not fit significantly worse than the model with factor loadings unconstrained across groups (χ2 (3, N = 1,817) = 1.78, p > .05). As such, academic aptitude was assumed to represent the same construct across groups and was included in structural models as a latent mediator.

The best-fitting, most parsimonious indirect multiple-group path model was identified through the creation of a series of models as described below (Keith, 2006; Sanchez, Whittaker, & Hamilton, 2016).

Model 1: single-group model with all G1 mothers

Model 2: single-group model with G1 teen mothers only

Model 3: single-group model with G1 non-teen mothers only

Model 4: multiple-group model with all parameters freed across groups

Model 5: multiple-group model with all parameters constrained across groups

Model 6: final indirect multiple-group model

Model 1.

A baseline single-group, model with all mother-child pairs (N = 1,817) was first estimated with G1 education variables regressed on all control variables and all hypothesized indirect paths between G1 education variables and G2 teenage childbearing (Figure 2). Fit indices for the baseline model indicated poor model fit as the baseline model assumed no direct paths between control variables and mediators, nor between children’s environmental and personal characteristics at age 14 and their probability of teenage childbearing. Informed by the modification indices, additional parameters were estimated, one by one, to modify the baseline model with all women until adequate model fit indices were achieved to arrive at Model 1 (Table 2).

Table 2.

Model Fit Indices

Model Description	RMSEA	CFI	TLI	WRMR
M1: Single group-all mothers	0.019	0.980	0.968	0.805
M2: Single group-teen mothers	0.024	0.988	0.980	0.681
M3: Single group-non-teen mothers	0.016	0.985	0.976	0.726
M4: Multiple group-all parameters free	0.018	0.984	0.975	1.020
M5: Multiple group-fully constrained^a	0.023	0.968	0.961	1.392
M6: Final multiple group^b	0.017	0.983	0.979	1.238

Open in a new tab

Model had significantly worse fit than model with all parameters freely estimated across both groups (M4) per results from adjusted chi-square difference tests for models using mean- and variance-adjusted weighted least squares (WLSMV) estimation, p < .01.

Model did not significantly reduce model fit when compared with the model with all parameters freely estimated across both groups (M4) per results from adjusted chi-square difference tests for models using mean- and variance-adjusted weighted least squares (WLSMV) estimation p > .05.

Note. GED: General Education Development/High School Equivalency Certificate; RMSEA: Root Mean Square Error of Approximation; CFI: Comparative Fit Index; TLI: Tucker Lewis Index; WRMR: Weighted Root Mean Square Residual (WRMR).All mothers (N = 1,817); teen mothers (n = 323); non-teen mothers (n = 1,494). All models controlled for Generation 1 (G1) age at Wave 1, G1 age at reporting educational outcomes, and scores for 2 principal components from a principal components analysis (PCA) of a polychoric correlation matrix of G1 participants’ baseline self-esteem scores, AFQT scores, highest grades expected, race/ethnicity, school enrollment status, number of siblings, and mothers’ educational attainment. Mediators included in all models were: environmental and personal characteristics when Generation 2 (G2) was age 14 (log of household income, perception of parental monitoring, academic aptitude) and G2 sexual risk behaviors in adolescence (age at first sex, and unprotected sex).

Model 2 & Model 3.

Model 1 was run separately with teen mothers only (Model 2) (n = 323) and non-teen mothers only (Model 3) (n = 1,494) to verify acceptable model fit within each group prior to conducting multiple group models (Sanchez et al., 2016). Fit indices indicated that the model fit the data sufficiently in each group (Table 2).

Model 4.

Model 1 was then estimated as a multiple-group model with no parameters constrained to be equal across the two groups (G1 teen mothers (n = 323) and G1 non-teen mothers (n = 1,494)). This model fit the data well (Table 2) and served as the baseline multiple-group model to which the more parsimonious, nested models (Model5 & Model 6) were compared.

Model 5.

Next, Model 4 was estimated with all parameters constrained to be equal across groups. This model represented the most parsimonious multiple-group model. Model 5 also fit the data well (Table 2), but constraining all parameters resulted in a significant loss in model fit when compared to Model 4: χ2 (29, N = 1,817) = 56.70, p < .01. This suggested that one or more parameters in the model significantly differed for teen mothers and non-teen mothers.

Model 6.

Finally, informed by the modification indices, parameters from Model 5 were unconstrained across groups until a model with adequate fit indices and a non-significant difference in model fit from Model 4 was achieved (χ2 (27, N = 1,817) = 32.87, p > .05) (Table 2). Model 6 served as the final indirect multiple-group model. Wald chi-squared tests of parameter equality were also conducted in the final multiple-group model to determine whether the magnitude of the relationships between mothers’ education and their children’s outcomes differed by the type of degree G1 mothers attained (GED v. high school diploma).

Results

The final multiple-group model (Model 6) is presented in Figure 4. Control variables are excluded from the figure for simplicity. The best-fitting, most parsimonious multiple-group model resulted in good fit indices for most indicators with the WRMR a little high (Table 2). The RMSEA, CFI, TLI and results from the χ2 difference test all suggested that the model fit the data well, and not significantly worse than the multiple-group model with all parameters freely estimated. Therefore, it was assumed that the model had adequate fit, and final results are drawn from this model. In addition to the paths presented in the conceptual model (Figure 2), the final model included a direct path from G2 log of household income to G2 teenage childbearing, a direct path from G1 high school diploma to G2 age at first sex, a correlation between G1 GED and high school diploma, and a correlation between G2 perception of parental monitoring and G2 academic performance (Figure 4). In addition to the paths depicted in Figure 4, the final model also included paths from all control variables to G1 degree types, and two additional paths from control variables (principal components factors #1 and #2) to G2 academic aptitude and G2 perception of parental monitoring, respectively.

Direct Relationships Moderated by Teen Motherhood Status

G1 high school diploma and G2 household income at age 14.

High school diploma attainment was significantly associated with higher household incomes for teen and non-teen mothers. However, non-teen mothers gained significantly more income for earning a high school diploma (compared to no degree) than teen mothers. Non-teen mothers who earned high school diplomas were expected to have household incomes approximately five times those of non-teen mothers who earned no degree, while teen mothers who earned high school diplomas were expected to have household incomes approximately four times those of teen mothers who earn no degree.

G1 high school diploma and G2 age at first sex.

Maternal high school diploma attainment was directly and significantly associated with a small increase in age (approximately .3 years) at first sex for the children of non-teen mothers above and beyond the association mediated through children’s environmental and personal characteristics at age 14 (Figure 4). High school diploma attainment was not directly associated with the age at first sex for the children of teen mothers.

Direct Relationships Not Moderated by Teen Motherhood Status

G1 GED and G2 household income at age 14.

Maternal GED attainment (compared to no degree) was associated with a fourfold increase in household income for both the children of teen and non-teen mothers. Wald test results indicated that high school diploma attainment was associated with a significantly larger increase (approximately 100% larger) in household incomes than GED attainment for non-teen mothers (χ² (1, N = 1,494) = 23.67, p < .001). In contrast, both teen mothers who earned GEDs and those who earned high school diplomas were expected to have household incomes approximately four times those of teen mothers who earned no degree (χ² (1, N = 323) = .13, p > .05).

G1 degree types and G2 perception of parental monitoring.

For children of both teen and non-teen mothers, maternal GED and high school diploma attainment were each associated with increases in the predicted probability of a child reporting high parental monitoring compared to the children of mothers who earned no degree. However, the R² for the probit model predicting high perceived parental monitoring was approximately .05 for both the children of teen and non-teen mothers. This suggests a small proportion (~5%) of the variance in the underlying continuous latent response variable associated with perceived parental monitoring for both groups was explained by the model. Results from the Wald test of parameter equality were nonsignificant, indicating that the difference between the gains in probability of perceiving high parental monitoring were not different for children of mothers who earned high school diplomas and the children of mothers who earned GEDs (χ² (1, N = 1,817) = 2.48, p > .05).

G1 degree types and G2 academic aptitude.

For children of both teen and non-teen mothers, G1 GED and high school diploma attainment were each associated with increases in G2 academic aptitude of approximately 3% in age-specific percentile scores. The difference between the increases in G2 academic aptitude by G1 degree type (high school diploma v. GED) was not statistically significant (χ² (1, N = 1,817) =0.22, p > .05).

G2 environmental and personal characteristics at age 14, sexual risk behaviors, and teenage childbearing.

Higher levels of academic aptitude at age 14 were associated with older ages at first sex and a reduced probability of reporting unprotected sex prior to age 19. Specifically, every 10% increase in age-specific academic aptitude percentile scores was associated with being approximately 1 year older at sexual debut. High perceived parental monitoring was associated with a slightly older age at first sex (approximately .25 years older) but was not associated with reporting unprotected sex prior to age 19. Household income when children were age 14 was not directly associated with either of the sexual risk behaviors in the final model. As expected, an older age at first sex was associated with a reduced probability of teenage childbearing. However, reporting unprotected sex before age 19 was not directly significantly associated with an increased probability of teenage childbearing in the final model. Also, higher household incomes at age 14 were directly associated with a significant reduction in probability of teenage childbearing for the children of teen and non-teen mothers, above and beyond the association mediated through sexual risk behaviors in adolescence. Together, sexual risk behaviors in adolescence and household income at age 14 explained 32% and 39% of the variance in the underlying continuous latent response variable associated with teenage childbearing for the children of teen and non-teen mothers, respectively.

Indirect Relationships Between G1 Degree Types and G2 Teenage Childbearing

Path coefficients for significant indirect relationships (at α = .05) between G1 degree types and G2 teenage childbearing are presented in Table 3 and represented with bold lines in Figure 4. All significant mediation pathways were the same for the children of both teen and non-teen mothers except one: youth’s age at first sex mediated the relationship between maternal high school diploma attainment and risk for teenage childbearing for the children of non-teen mothers only.

Table 3.

Significant indirect paths from G1 degree type attained by age 30 to G2 teenage childbearing

				Teen mothers n = 323			Non-teen mothers n = 1,494
G1 degree type (v. no degree)	G2 characteristic at age 14	G2 sexual risk behavior in adolescence	G2 teenage childbearing	B	SE B		B	SE B
GED →	parental monitoring →	age at first sex →	first birth before age 19	−.059	.029	^*	−.059	.029	^*
GED →	academic aptitude →	age at first sex →	first birth before age 19	−.129	.050	^*	−.129	.050	^*
GED →	household income →		first birth before age 19	−.439	.111	^***	−.439	.111	^***
HS →	parental monitoring →	age at first sex →	first birth before age 19	−.067	.030	^*	−.067	.030	^*
HS →	academic aptitude →	age at first sex →	first birth before age 19	−.131	.050	^*	−.131	.050	^*
HS →		age at first sex →	first birth before age 19	.040	.041		−.126	.033	^***
HS →	household income →		first birth before age 19	−.437	.112	^***	−.495	.120	^***

Open in a new tab

p <.05

^**

p <.01

^***

p <.001.

Note. Unstandardized indirect path coefficients and standard errors shown. GED: General Education Development/High School Equivalency Certificate. HS: high school diploma. G1: Generation #1. G2: Generation #2. Bolded indirect paths were freely estimated across groups in final multiple-group model.

Discussion

The current study aimed to determine whether the mechanisms underlying the relationship between mothers’ high school degree type and children’s risk for teenage childbearing differ by the mothers’ teen motherhood status using data from a longitudinal, multigenerational sample.

We hypothesized that higher levels of education for teen mothers would be associated with a reduced risk of their children experiencing teenage childbearing. Based on social ecological models and previous literature noting the causes and consequences of children’s academic aptitude (Lou & Thomas, 2015; Tang et al., 2014), experiences of parental monitoring(Huebner & Howell, 2003), and household incomes (Carlson et al., 2014; US Bureau of Labor Statistics, 2014), we posited that these would each partially mediate the association between teen mothers’ educational attainment and their children’s risk for sexual behaviors in adolescence and risk for teenage childbearing. Drawing from the resource substitution theory (Ross & Mirowsky, 2006), we predicted the associations of maternal educational attainment with later-life outcomes would be greater for teen mothers than non-teen mothers. We further examined whether maternal high school degree attainment was associated with a larger association with children’s outcomes than GED attainment when compared to earning no degree for both teen and non-teen mothers.

Mechanisms by Which Maternal Education Level Predicts Risk for Teenage Childbearing

Supporting our hypothesis, and consistent with previous literature, we found maternal education level to predict children’s personal and environmental contexts in adolescence (Magnuson, 2007; Tang et al., 2014), which, in turn, predicted adolescents’ sexual risk behaviors and risk for teenage childbearing (Carlson et al., 2014; Huebner & Howell, 2003; Kirby & Lepore, 2007). As predicted, higher levels of children’s perceptions of parental monitoring and academic aptitude in adolescence were each associated with older ages at first sex.

However, contrary to our hypotheses and previous literature (DiClemente et al., 2001; Dittus et al., 2015; Li et al., 2000; Miller, 2002), youth’s perceptions of parental monitoring were not directly associated with their probability for reporting unprotected sex, and household incomes in adolescence were not significantly associated with either of the measured sexual risk behaviors in adolescence. As household income was directly associated with probability for teenage childbearing in the final model, this may partially explain why the relationships between household income and sexual risk behaviors in adolescence were no longer significant. Further, past research has demonstrated parental monitoring to predict some sexual risk behaviors but not others (Crosby, Terrell, & Pasternak, 2015; Romer et al., 1999). For example, among a sample of African American adolescents living in high-poverty urban areas, Romer and colleagues found higher levels of perceived parental monitoring to be associated with later sexual debut, but not associated with condom use (1999). Thus, additional research is needed for a more comprehensive examination of how youth’s perceived parental monitoring predicts combinations of correlated adolescent sexual risk behaviors.

Most mechanisms were similar for teen and non-teen mothers.

Our findings also suggest that the mechanisms by which women’s educational attainment leads to long-term health and socioeconomic outcomes for women and children are similar for teen and non-teen mothers. Specifically, six of the seven significant indirect pathways between higher maternal education level and children’s reduced risk for teenage childbearing for non-teen mothers were also significant for teen mothers. This finding is notable because it suggests that teen mothers and their offspring experience similar associations, and a similar magnitude of associations, of maternal educational attainment with later-life sequelae as non-teen mothers. This finding is consistent with some findings of Meade and colleagues (2008) in that they found maternal education level and school performance to significantly predict teenage childbearing among the daughters of both teen and non-teen mothers in a sample of women from the NLSY97 cohort. However, they found low parental monitoring and living below the poverty threshold to predict teenage childbearing for only the daughters of teen mothers.

Our findings add to those of Meade and colleagues in important ways. First, we include both sons and daughters as both sons and daughters of teen mothers are at elevated risk for teenage childbearing (Kim, 2014). Also, due to the use of data from a multigenerational cohort, we were able to rigorously control for background characteristics in mothers’ lives that may confound the associations of maternal and household characteristics with youth’s risk for teenage childbearing. We also defined maternal education level by degree type (no degree, GED, or high school diploma) whereas Meade and colleagues defined maternal education as completing ≥12 years or <12 years. Since teen mothers and non-teen mothers earn GEDs and high school diplomas at dissimilar rates (Perper et al., 2010), this distinction is important when assessing the relationship of maternal education with outcomes for the children of teen versus non-teen mothers. Thus, in addition to examining the direct relationships of maternal education and personal and environmental characteristics with youths’ risk for teenage childbearing, we elucidated several mediating pathways of these relationships to better inform our understanding of the development of risk for teenage childbearing, as well as potential points of intervention along the life course.

Few mechanisms were moderated by teen motherhood status.

Although most relationships in the current study were not moderated by teen motherhood status, the relationship between maternal high school diploma attainment and household income when children were approximately 14 years of age, as well as the relationship between maternal high school diploma attainment and youths’ age at sexual debut, were significantly moderated by teen motherhood status. Mothers earning high school diplomas were expected to have household incomes approximately five and four times those of mothers with no degree for non-teen and teen mothers, respectively. This finding did not support our hypothesis based on the resource substitution theory (Ross & Mirowsky, 2006) and related literature finding younger mothers from lower-resource settings to experience greater returns from increases in educational attainment (Magnuson, 2007). If there are differences in the associations of educational attainment with later-life outcomes by age at childbearing, findings from the current study suggest they are in favor of women who delay childbearing beyond the teenage years. Before policy and programming recommendations can be made regarding these findings, it needs to be determined whether the lower gains associated with teen mothers’ educational attainment are due to qualitative differences in the educational experiences of teen and non-teen mothers, or to unmeasured characteristics that influence how teen and older mothers’ high school degree impact their and their children’s later life sequalae. The challenges teen mothers experience when attempting to continue their education during pregnancy and after their children are born (Kalil, 2002; Kaufman & Malone, 2015; SmithBattle, 2007) imply that teen mothers may be having different educational experiences, which may partially explain findings of differential associations of educational attainment with household income for teen and non-teen mothers. If this is the case, important policy implications would be to improve the school environment for pregnant and parenting students to both 1) address their unique needs as student parents, and 2) ensure they are afforded the same educational opportunities as non-pregnant/parenting students.

Household income directly predicted risk for teenage childbearing.

A notable, unpredicted finding was that, for both the children of teen and non-teen mothers, higher household incomes were associated with a direct reduction in the probability of teenage childbearing, above and beyond that explained through youth’s sexual risk behaviors during the teenage years. This may signify that there are additional mediating variables we were not able to include in our model. For example, youth from higher-resource settings may have more access to more expensive, more effective contraception methods (e.g., intrauterine devices and subdermal implants) than youth from lower-resource settings and consequently have lower risk for teenage childbearing even when demonstrating the same sexual risk behaviors during the teenage years. Similarly, as teenage childbearing, not teenage pregnancy, was assessed as the outcome of interest in the current study, it is possible that youth from higher-resource settings have more access to abortion care if an unwanted pregnancy occurs during the teenage years. This finding supports previous research noting youth from higher-resource settings to have more access to reproductive health care services (Hall et al., 2012).

The GED and the High School Diploma for Teen and Non-teen Mothers

Findings were mixed regarding whether mothers’ high school diplomas had greater associations with later-life outcomes than GEDs when compared to earning no degree. The association of earning a high school diploma (versus a GED) with household incomes was greater for non-teen mothers; however, there was no significant difference for teen mothers. Maternal high school diploma attainment did not differ from the GED in its strength of association with children’s perceived parental monitoring, nor academic aptitude at age 14, for the children of either group of mothers. As teen mothers earn GEDs at elevated rates (Perper et al., 2010), this is an important area of inquiry for improving our understanding of the relationship between teen mothers’ educational experiences and their and their children’s long-term outcomes. However, considering these mixed findings, teen mothers’ outcomes associated with earning a GED should continue to be studied in future research.

Limitations

It is important to interpret and translate the findings from this study to the experiences of current teen mothers with caution due to potential cohort effects. While longitudinal, and therefore older, data were necessary to examine the current research questions, it is important to remember that women in the analytic sample started childbearing in the 1970’s and completed their education in the 70’s and 80’s. The United States has experienced dramatic changes in teen birth rates, characteristics of teen mothers, and women’s educational attainment since the 1970’s (U.S. Census Bureau, 2015; Ventura, Hamilton, & Matthews, 2014). Further, the small sample size of teen mothers must also be considered when interpreting the findings from the current study. While we included several potentially confounding covariates to address observed selection factors, results cannot be interpreted as causal. It is possible that unobserved variables explain findings. Specifically, although a sensitivity analysis was conducted controlling for a measure representing childhood risk behaviors that did not yield statistically, nor substantively different results, it is still possible that results reflect stability of risk behavior from childhood to adolescence rather than a causal pathway among observed variables. Further, with 323 teen mothers in the analytic sample and the number of parameters estimated in models, we were unable to address confounding effects of control variables individually, but rather collectively represented as principal component scores. It is also possible that additional significant moderation effects by teen motherhood status may have been detected with a larger sample of teen mothers. Also due to the small sample of teen mothers, it was not possible to further examine whether the mechanisms by which teen mothers’ educational attainment was associated with their children’s probability of teenage childbearing varied by child’s gender nor race/ethnicity. Relatedly, because the number of teen mothers completing postsecondary education in the analytic sample was so small, it was not possible to assess postsecondary education as a predictor in the current study. Finally, these analyses should be interpreted in light of a common challenge in studying low-frequency outcomes, which is that the subgroup of children of teen mothers who experienced the birth of their first child prior to age 19 was small (n = 45, 14% of the 323 children of teen mothers). Future studies with larger samples may identify additional predictive pathways to childbearing among children of teen mothers.

Directions for Future Research

Due to the overall increases in women’s educational attainment in the US, the increases in single-mother led households, and the minimum education requirements for earning a living wage present day, it is possible that the benefits once seen from mothers attaining a high school diploma or a GED in the 1970’s and 80’s may now only be seen with the completion of post-secondary education. This may be especially true for teen mothers as the percentage of unmarried teen mothers has increased dramatically from the 1970’s to 2013, from 29.5% to 88.7% (Ventura et al., 2014). Future studies should continue to assess the later-life outcomes associated with teen mothers’ educational attainment in more recent cohorts and examine whether post-secondary degree attainment has become necessary for significant long-term outcomes for teen mothers and their children.

Previous research has noted different mechanisms by which teenage childbearing is transmitted from mothers to sons and daughters (Barber, 2001; Wildsmith et al., 2012). Future studies that specifically examine the role of teen and non-teen mothers’ educational attainment as it relates to their sons’ and daughters’ risk for teenage childbearing would be an important contribution to this field. Similarly, further examining the role of maternal educational attainment in the intergenerational cycle of teenage childbearing within and across racial and ethnic groups is warranted given disparities in women’s educational experiences and teenage childbearing rates across racial and ethnic lines in the United States (Martin et al., 2018).

Conclusions

Despite the additional challenges of raising children as young mothers and the background environmental and personal characteristics that selected women into early childbearing, teen mothers in this sample experienced long-term positive associations with their educational attainment, and with later-life mediators related to their education, on their children’s risk for teenage childbearing. Furthermore, these long-term outcomes for teen mothers were mostly similar to those of non-teen mothers. As such, the findings from the current study provide support for programming that aims to assist teen mothers in attaining higher levels of education as higher levels of education are associated with positive later-life sequelae for both teen mothers and their children.

As the children of teen mothers continue to be at elevated risk for teenage childbearing, and teenage childbearing is associated with a cycle of poverty and poor health for families, it is important to identify the salient, modifiable mechanisms for reducing the intergenerational cycle of teenage childbearing. The findings from this study highlight the potential of teen mothers’ educational attainment as a strategy for reducing their children’s risk for teenage childbearing, but also point to a need for additional research in this area to provide a more comprehensive understanding of the relationship between teen mothers’ educational attainment and their children’s risk for teenage childbearing.

Acknowledgments

This research received support from grants from the Eunice Kennedy Shriver National Institute of Child Health and Human Development: T32HD007081, R24HD042849, T32HD049302, and K01HD091416. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Contributor Information

C. Emily Hendrick, Division of Reproduction and Population Health, Department of Obstetrics and Gynecology, School of Medicine and Public Health, University of Wisconsin-Madison.

Julie Maslowsky, Department of Kinesiology and Health Education, College of Education, The University of Texas at Austin.

References

Assini-Meytin LC, & Green KM (2015). Long-term consequences of adolescent parenthood among African-American urban youth: A propensity score matching approach. Journal of Adolescent Health, 56(5), 529–535. 10.1016/j.jadohealth.2015.01.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
Barber JS (2001). The intergenerational transmission of age at first birth among married and unmarried men and women. Social Science Research, 30(2), 219–247. 10.1006/ssre.2000.0697 [DOI] [Google Scholar]
Bauldry S (2015). Variation in the protective effect of higher education against depression. Society and Mental Health, 5(2), 145–161. 10.1177/2156869314564399 [DOI] [PMC free article] [PubMed] [Google Scholar]
Cameron SV, & Heckman JJ (1993). The nonequivalence of high school equivalents. Journal of Labor Economics, 11(1, Part 1), 1–47. 10.1086/298316 [DOI] [Google Scholar]
Campa MI, & Eckenrode JJ (2006). Pathways to intergenerational adolescent childbearing in a high-risk sample. Journal of Marriage and Family, 68(3), 558–572. 10.1111/j.1741-3737.2006.00275.x [DOI] [Google Scholar]
Carlson DL, McNulty TL, Bellair PE, & Watts S (2014). Neighborhoods and racial/ethnic disparities in adolescent sexual risk behavior. Journal of Youth and Adolescence, 43(9), 1536–1549. 10.1007/s10964-013-0052-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
Carneiro P, Meghir C, & Parey M (2013). Maternal education, home environments, and the development of children and adolescents. Journal of the European Economic Association, 11(S1), 123–160. 10.1111/j.1542-4774.2012.01096.x [DOI] [Google Scholar]
Coley SL, Nichols TR, Rulison KL, Aronson RE, Brown-Jeffy SL, & Morrison SD (2016). Does neighborhood risk explain racial disparities in low birth weight among infants born to adolescent mothers? Journal of Pediatric and Adolescent Gynecology, 29(2), 122–129. 10.1016/j.jpag.2015.08.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
Costello CB, & Institute for Women’s Policy Research (IWPR). (2014, May). Pathways to postsecondary education for pregnant and parenting teens Working paper #C418. Institute for Women’s Policy Research; Retrieved from http://files.eric.ed.gov/fulltext/ED556724.pdf [Google Scholar]
Crosby R, Terrell I, & Pasternak R (2015). Is perceived parental monitoring associated with sexual risk behaviors of young Black males? Preventive Medicine Reports, 2, 829–832. 10.1016/j.pmedr.2015.09.014 [DOI] [PMC free article] [PubMed] [Google Scholar]
Crosnoe R, Ansari A, Purtell KM, & Wu N (2016). Latin American immigration, maternal education, and approaches to managing children’s schooling in the United States. Journal of Marriage and the Family, 78(1), 60–74. 10.1111/jomf.12250 [DOI] [PMC free article] [PubMed] [Google Scholar]
Davis-Kean PE (2005). The influence of parent education and family income on child achievement: The indirect role of parental expectations and the home environment. Journal of Family Psychology, 19(2), 294–304. 10.1037/0893-3200.19.2.294 [DOI] [PubMed] [Google Scholar]
DiClemente RJ, Salazar LF, Crosby RA, & Rosenthal SL (2005). Prevention and control of sexually transmitted infections among adolescents: The importance of a socio-ecological perspective—A commentary. Public Health, 119(9), 825–836. 10.1016/j.puhe.2004.10.015 [DOI] [PubMed] [Google Scholar]
DiClemente RJ, Wingood GM, Crosby R, Sionean C, Cobb BK, Harrington K, … Oh MK (2001). Parental monitoring: Association with adolescents’ risk behaviors. Pediatrics, 107(6), 1363–1368. 10.1542/peds.107.6.1363 [DOI] [PubMed] [Google Scholar]
Dittus PJ, Michael SL, Becasen JS, Gloppen KM, McCarthy K, & Guilamo-Ramos V (2015). Parental monitoring and its associations with adolescent sexual risk behavior: A meta-analysis. Pediatrics, 136(6), e1587–1599. 10.1542/peds.2015-0305 [DOI] [PMC free article] [PubMed] [Google Scholar]
Domina T, & Roksa J (2012). Should Mom go back to school? Post-natal educational attainment and parenting practices. Social Science Research, 41(3), 695–708. 10.1016/j.ssresearch.2011.12.002 [DOI] [PubMed] [Google Scholar]
East PL, Reyes BT, & Horn EJ (2007). Association between adolescent pregnancy and a family history of teenage births. Perspectives on Sexual and Reproductive Health, 39(2), 108–115. 10.1363/3910807 [DOI] [PMC free article] [PubMed] [Google Scholar]
Fry R, & Taylor P (2012, August 1). The rise of residential segregation by income Retrieved from http://www.pewsocialtrends.org/2012/08/01/the-rise-of-residential-segregation-by-income/
Gibb SJ, Fergusson DM, Horwood LJ, & Boden JM (2015). Early motherhood and long-term economic outcomes: Findings from a 30-year longitudinal study. Journal of Research on Adolescence, 25(1), 163–172. 10.1111/jora.12122 [DOI] [Google Scholar]
Graefe DR, & Lichter DT (2002). Marriage among unwed mothers: Whites, Blacks and Hispanics compared. Perspectives on Sexual and Reproductive Health, 34(6), 286–293. 10.2307/3097747 [DOI] [PubMed] [Google Scholar]
Greenwood J, Guner N, Kocharkov G, & Santos C (2014). Marry your like: Assortative mating and income inequality. The American Economic Review, 104(5), 348–353. 10.1257/aer.104.5.348 [DOI] [Google Scholar]
Hall KS, Moreau C, & Trussell J (2012). Determinants of and disparities in reproductive health service use among adolescent and young adult women in the United States, 2002–2008. American Journal of Public Health, 102(2), 359–367. 10.2105/AJPH.2011.300380 [DOI] [PMC free article] [PubMed] [Google Scholar]
Harding JF (2015). Increases in maternal education and low-income children’s cognitive and behavioral outcomes. Developmental Psychology, 51(5), 583–599. http://dx.doi.org.ezproxy.library.wisc.edu/10.1037/a0038920 [DOI] [PubMed] [Google Scholar]
Harding JF, Morris PA, & Hughes D (2015). The relationship between maternal education and children’s academic outcomes: A theoretical framework. Journal of Marriage and Family, 77(1), 60–76. [Google Scholar]
Henretta JC (2007). Early childbearing, marital status, and women’s health and mortality after age 50. Journal of Health and Social Behavior, 48(3), 254–266. 10.1177/002214650704800304 [DOI] [PubMed] [Google Scholar]
Hu L, & Bentler PM (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6(1), 1–55. 10.1080/10705519909540118 [DOI] [Google Scholar]
Huebner AJ, & Howell LW (2003). Examining the relationship between adolescent sexual risk-taking and perceptions of monitoring, communication, and parenting styles. Journal of Adolescent Health, 33(2), 71–78. 10.1016/S1054-139X(03)00141-1 [DOI] [PubMed] [Google Scholar]
Jaffee S, Caspi A, Moffitt TE, Belsky J, & Silva P (2001). Why are children born to teen mothers at risk for adverse outcomes in young adulthood? Results from a 20-year longitudinal study. Development and Psychopathology, 13(2), 377–397. 10.1017/S0954579401002103 [DOI] [PubMed] [Google Scholar]
Jutte DP, Roos NP, Brownell MD, Briggs G, MacWilliam L, & Roos LL (2010). The ripples of adolescent motherhood: Social, educational, and medical outcomes for children of teen and prior teen mothers. Academic Pediatrics, 10(5), 293–301. 10.1016/j.acap.2010.06.008 [DOI] [PubMed] [Google Scholar]
Kane JB, Philip Morgan S, Harris KM, & Guilkey DK (2013). The educational consequences of teen childbearing. Demography, 50(6), 2129–2150. 10.1007/s13524-013-0238-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
Keith TZ (2006). Multisample Models. In Multiple Regression and Beyond (1st ed., pp. 362–374). Boston, MA: Pearson Education, Inc. [Google Scholar]
Kim K (2014). Intergenerational transmission of age at first birth in the United States: Evidence from multiple surveys. Population Research and Policy Review, 33(5), 649–671. 10.1007/s11113-014-9328-7 [DOI] [Google Scholar]
Kirby D, & Lepore G (2007, November 26). Sexual risk and protective factors: Factors affecting teen sexual behavior, pregnancy, childbearing and sexually transmitted disease: Which are important? Which can you change? ETR Associates; Retrieved from http://recapp.etr.org/recapp/documents/theories/RiskProtectiveFactors200712.pdf [Google Scholar]
Kolenikov S, & Angeles G (2009). Socioeconomic status measurement with discrete proxy variables: Is principal component analysis a reliable answer? Review of Income and Wealth, 55(1), 128–165. 10.1111/j.1475-4991.2008.00309.x [DOI] [Google Scholar]
Lee D (2010). The early socioeconomic effects of teenage childbearing: A propensity score matching approach. Demographic Research, 23, 697–736. 10.4054/DemRes.2010.23.25 [DOI] [Google Scholar]
Lee Y (2009). Early motherhood and harsh parenting: The role of human, social, and cultural capital. Child Abuse & Neglect, 33(9), 625–637. 10.1016/j.chiabu.2009.02.007 [DOI] [PubMed] [Google Scholar]
Lewin A, Mitchell SJ, & Ronzio CR (2013). Developmental differences in parenting behavior: Comparing adolescent, emerging adult, and adult mothers. Merrill-Palmer Quarterly, 59(1), 23–49. 10.1353/mpq.2013.0003 [DOI] [Google Scholar]
Li X, Stanton B, & Feigelman S (2000). Impact of perceived parental monitoring on adolescent risk behavior over 4 years. Journal of Adolescent Health, 27(1), 49–56. 10.1016/S1054-139X(00)00092-6 [DOI] [PubMed] [Google Scholar]
Lindberg L, Santelli J, & Desai S (2016). Understanding the decline in adolescent fertility in the United States, 2007–2012. Journal of Adolescent Health, 59(5), 577–583. 10.1016/j.jadohealth.2016.06.024 [DOI] [PMC free article] [PubMed] [Google Scholar]
Lou C, & Thomas A (2015). The relationship between academic achievement and nonmarital teenage childbearing: Evidence from the panel study of income dynamics. Perspectives on Sexual and Reproductive Health, 47(2), 91–98. 10.1363/47e2115 [DOI] [PubMed] [Google Scholar]
Luster T, & Small SA (1994). Factors associated with sexual risk-taking behaviors among adolescents. Journal of Marriage and the Family, 56(3), 622–632. 10.2307/352873 [DOI] [Google Scholar]
Magnuson K (2007). Maternal education and children’s academic achievement during middle childhood. Developmental Psychology, 43(6), 1497–1512. 10.1037/0012-1649.43.6.1497 [DOI] [PubMed] [Google Scholar]
Manlove J (1998). The influence of high school dropout and school disengagement on the risk of school-age pregnancy. Journal of Research on Adolescence, 8(2), 187–220. 10.1207/s15327795jra0802_2 [DOI] [PubMed] [Google Scholar]
Markwardt FC (1989). Peabody individual achievement test-revised Circle Pines, MN: American Guidance Service. [Google Scholar]
Martin JA, Hamilton BE, Osterman MJK, Driscoll AK, & Drake P (2018, January 31). Births: Final data for 2016 (National Vital Statistics Reports Vol. 67 No. 1). National Center for Health Statistics; Retrieved from https://www.cdc.gov/nchs/data/nvsr/nvsr67/nvsr67_01.pdf [PubMed] [Google Scholar]
Meade CS, Kershaw TS, & Ickovics JR (2008). The intergenerational cycle of teenage motherhood: An ecological approach. Health Psychology, 27(4), 419 10.1037/0278-6133.27.4.419 [DOI] [PubMed] [Google Scholar]
Miller BC (2002). Family influences on adolescent sexual and contraceptive behavior. Journal of Sex Research, 39(1), 22–26. 10.1080/00224490209552115 [DOI] [PubMed] [Google Scholar]
Muthén LK, & Muthén BO (2012). Mplus user’s guide seventh edition. Los Angeles, CA:Muthén & Muthén. [Google Scholar]
Perper K, Peterson K, & Manlove J (2010). Diploma attainment among teen mothers. Child Trends Retrieved from https://www.childtrends.org/wp-content/uploads/2010/01/child_trends-2010_01_22_FS_diplomaattainment.pdf
Pillow WS (2004). Unfit subjects: Educational policy and the teen mother New York, NY, US: RoutledgeFalmer. [Google Scholar]
Pogarsky G, Thornberry TP, & Lizotte AJ (2006). Developmental outcomes for children of young mothers. Journal of Marriage and Family, 68(2), 332–344. 10.1111/j.1741-3737.2006.00256.x [DOI] [Google Scholar]
Rai AA, Stanton B, Wu Y, Li X, Galbraith J, Cottrell L, … Burns J (2003). Relative influences of perceived parental monitoring and perceived peer involvement on adolescent risk behaviors: An analysis of six cross-sectional data sets. Journal of Adolescent Health, 33(2), 108–118. 10.1016/s1054-139x(03)00179-4 [DOI] [PubMed] [Google Scholar]
Romer D, Stanton B, Galbraith J, Feigelman S, Black MM, & Li X (1999). Parental influence on adolescent sexual behavior in high-poverty settings. Archives of Pediatrics & Adolescent Medicine, 153(10), 1055–1062. 10.1001/archpedi.153.10.1055 [DOI] [PubMed] [Google Scholar]
Ross CE, & Mirowsky J (2006). Sex differences in the effect of education on depression: resource multiplication or resource substitution? Social Science & Medicine, 63(5), 1400–1413. 10.1016/j.socscimed.2006.03.013 [DOI] [PubMed] [Google Scholar]
Ross CE, & Mirowsky J (2011). The interaction of personal and parental education on health. Social Science & Medicine, 72(4), 591–599. 10.1016/j.socscimed.2010.11.028 [DOI] [PMC free article] [PubMed] [Google Scholar]
Sanchez D, Whittaker TA, & Hamilton E (2016). Perceived discrimination, peer influence and sexual behaviors in Mexican American preadolescents. Journal of Youth and Adolescence, 45(5), 928–944. 10.1007/s10964-016-0420-7 [DOI] [PubMed] [Google Scholar]
Schreiber JB, Nora A, Stage FK, Barlow EA, & King J (2006). Reporting structural equation modeling and confirmatory factor analysis results: A review. The Journal of Educational Research, 99(6), 323–338. 10.3200/joer.99.6.323-338 [DOI] [Google Scholar]
Strickland BB, Jones JR, Ghandour RM, Kogan MD, & Newacheck PW (2011). The medical home: Health care access and impact for children and youth in the United States. Pediatrics, 127(4), 604–611. 10.1542/peds.2009-3555 [DOI] [PubMed] [Google Scholar]
Sullivan CJ, Childs KK, & O’Connell D (2010). Adolescent risk behavior subgroups: An empirical assessment. Journal of Youth and Adolescence, 39(5), 541–562. 10.1007/s10964-009-9445-5 [DOI] [PubMed] [Google Scholar]
Tang S, Davis-Kean PE, Chen M, & Sexton HR (2014). Adolescent pregnancy’s intergenerational effects: Does an adolescent mother’s education have consequences for her children’s achievement? Journal of Research on Adolescence, 26(1), 180–193. 10.1111/jora.12182 [DOI] [Google Scholar]
US Bureau of Labor Statistics. (2014, June 5). Educational attainment and earnings of women Retrieved from http://www.bls.gov/opub/ted/2014/ted_20140603.htm
US Bureau of Labor Statistics. (n.d.-a). National Longitudinal Survey of Youth 1979 sample weights & clustering adjustments: Practical usage of weights Retrieved from https://www.nlsinfo.org/content/cohorts/nlsy79/using-and-understanding-the-data/sample-weights-clustering-adjustments
US Bureau of Labor Statistics. (n.d.-b). The NLSY79 Retrieved from https://www.bls.gov/nls/nlsy79.htm
US Bureau of Labor Statistics. (n.d.-c). The NLSY79 Child/Young Adult sample: An introduction Retrieved from https://www.nlsinfo.org/content/cohorts/nlsy79-children/intro-to-the-sample/nlsy79-childyoung-adult-sample-introduction
US Bureau of Labor Statistics. (n.d.-d). National Longitudinal Survey of Youth 1979 Children and Young Adults: Behavior Problem Index (BPI) Retrieved from https://www.nlsinfo.org/content/cohorts/nlsy79-children/topical-guide/assessments/behavior-problems-index-bpi
U.S. Census Bureau. (2015, June). Annual estimates of the resident population by sex, single year of age, race, and Hispanic origin for the United States: April 1, 2010 to July 1, 2014 Retrieved from http://factfinder.census.gov/bkmk/table/1.0/en/PEP/2014/PEPALL6N?slice=hisp~nhisp
Ventura SJ, Hamilton BE, & Matthews TJ (2014, August 20). National and state patterns of teen births in the United States, 1940–2013 (National Vital Statistics Reports Vol. 63 No. 4). National Center for Health Statistics; Retrieved from https://www.cdc.gov/nchs/data/nvsr/nvsr63/nvsr63_04.pdf [PubMed] [Google Scholar]
Wildsmith E, Manlove J, Jekielek S, Moore KA, & Mincieli L (2012). Teenage childbearing among youth born to teenage mothers. Youth & Society, 44(2), 258–283. 10.1177/0044118X11398366 [DOI] [PMC free article] [PubMed] [Google Scholar]
Yu CY (2002). Evaluating cutoff criteria of model fit indices for latent variable models with binary and continuous outcomes Doctoral dissertation, University of California Los Angeles; Retrieved from https://pdfs.semanticscholar.org/7a22/ae22553f78582fc61c6cab4567d36998293b.pdf [Google Scholar]
Zajacova A, & Everett BG (2014). The nonequivalent health of high school equivalents. Social Science Quarterly, 95(1), 221–238. 10.1111/ssqu.12039 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R1] Assini-Meytin LC, & Green KM (2015). Long-term consequences of adolescent parenthood among African-American urban youth: A propensity score matching approach. Journal of Adolescent Health, 56(5), 529–535. 10.1016/j.jadohealth.2015.01.005 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R2] Barber JS (2001). The intergenerational transmission of age at first birth among married and unmarried men and women. Social Science Research, 30(2), 219–247. 10.1006/ssre.2000.0697 [DOI] [Google Scholar]

[R3] Bauldry S (2015). Variation in the protective effect of higher education against depression. Society and Mental Health, 5(2), 145–161. 10.1177/2156869314564399 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R4] Cameron SV, & Heckman JJ (1993). The nonequivalence of high school equivalents. Journal of Labor Economics, 11(1, Part 1), 1–47. 10.1086/298316 [DOI] [Google Scholar]

[R5] Campa MI, & Eckenrode JJ (2006). Pathways to intergenerational adolescent childbearing in a high-risk sample. Journal of Marriage and Family, 68(3), 558–572. 10.1111/j.1741-3737.2006.00275.x [DOI] [Google Scholar]

[R6] Carlson DL, McNulty TL, Bellair PE, & Watts S (2014). Neighborhoods and racial/ethnic disparities in adolescent sexual risk behavior. Journal of Youth and Adolescence, 43(9), 1536–1549. 10.1007/s10964-013-0052-0 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R7] Carneiro P, Meghir C, & Parey M (2013). Maternal education, home environments, and the development of children and adolescents. Journal of the European Economic Association, 11(S1), 123–160. 10.1111/j.1542-4774.2012.01096.x [DOI] [Google Scholar]

[R8] Coley SL, Nichols TR, Rulison KL, Aronson RE, Brown-Jeffy SL, & Morrison SD (2016). Does neighborhood risk explain racial disparities in low birth weight among infants born to adolescent mothers? Journal of Pediatric and Adolescent Gynecology, 29(2), 122–129. 10.1016/j.jpag.2015.08.004 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R9] Costello CB, & Institute for Women’s Policy Research (IWPR). (2014, May). Pathways to postsecondary education for pregnant and parenting teens Working paper #C418. Institute for Women’s Policy Research; Retrieved from http://files.eric.ed.gov/fulltext/ED556724.pdf [Google Scholar]

[R10] Crosby R, Terrell I, & Pasternak R (2015). Is perceived parental monitoring associated with sexual risk behaviors of young Black males? Preventive Medicine Reports, 2, 829–832. 10.1016/j.pmedr.2015.09.014 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R11] Crosnoe R, Ansari A, Purtell KM, & Wu N (2016). Latin American immigration, maternal education, and approaches to managing children’s schooling in the United States. Journal of Marriage and the Family, 78(1), 60–74. 10.1111/jomf.12250 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R12] Davis-Kean PE (2005). The influence of parent education and family income on child achievement: The indirect role of parental expectations and the home environment. Journal of Family Psychology, 19(2), 294–304. 10.1037/0893-3200.19.2.294 [DOI] [PubMed] [Google Scholar]

[R13] DiClemente RJ, Salazar LF, Crosby RA, & Rosenthal SL (2005). Prevention and control of sexually transmitted infections among adolescents: The importance of a socio-ecological perspective—A commentary. Public Health, 119(9), 825–836. 10.1016/j.puhe.2004.10.015 [DOI] [PubMed] [Google Scholar]

[R14] DiClemente RJ, Wingood GM, Crosby R, Sionean C, Cobb BK, Harrington K, … Oh MK (2001). Parental monitoring: Association with adolescents’ risk behaviors. Pediatrics, 107(6), 1363–1368. 10.1542/peds.107.6.1363 [DOI] [PubMed] [Google Scholar]

[R15] Dittus PJ, Michael SL, Becasen JS, Gloppen KM, McCarthy K, & Guilamo-Ramos V (2015). Parental monitoring and its associations with adolescent sexual risk behavior: A meta-analysis. Pediatrics, 136(6), e1587–1599. 10.1542/peds.2015-0305 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R16] Domina T, & Roksa J (2012). Should Mom go back to school? Post-natal educational attainment and parenting practices. Social Science Research, 41(3), 695–708. 10.1016/j.ssresearch.2011.12.002 [DOI] [PubMed] [Google Scholar]

[R17] East PL, Reyes BT, & Horn EJ (2007). Association between adolescent pregnancy and a family history of teenage births. Perspectives on Sexual and Reproductive Health, 39(2), 108–115. 10.1363/3910807 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R18] Fry R, & Taylor P (2012, August 1). The rise of residential segregation by income Retrieved from http://www.pewsocialtrends.org/2012/08/01/the-rise-of-residential-segregation-by-income/

[R19] Gibb SJ, Fergusson DM, Horwood LJ, & Boden JM (2015). Early motherhood and long-term economic outcomes: Findings from a 30-year longitudinal study. Journal of Research on Adolescence, 25(1), 163–172. 10.1111/jora.12122 [DOI] [Google Scholar]

[R20] Graefe DR, & Lichter DT (2002). Marriage among unwed mothers: Whites, Blacks and Hispanics compared. Perspectives on Sexual and Reproductive Health, 34(6), 286–293. 10.2307/3097747 [DOI] [PubMed] [Google Scholar]

[R21] Greenwood J, Guner N, Kocharkov G, & Santos C (2014). Marry your like: Assortative mating and income inequality. The American Economic Review, 104(5), 348–353. 10.1257/aer.104.5.348 [DOI] [Google Scholar]

[R22] Hall KS, Moreau C, & Trussell J (2012). Determinants of and disparities in reproductive health service use among adolescent and young adult women in the United States, 2002–2008. American Journal of Public Health, 102(2), 359–367. 10.2105/AJPH.2011.300380 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R23] Harding JF (2015). Increases in maternal education and low-income children’s cognitive and behavioral outcomes. Developmental Psychology, 51(5), 583–599. http://dx.doi.org.ezproxy.library.wisc.edu/10.1037/a0038920 [DOI] [PubMed] [Google Scholar]

[R24] Harding JF, Morris PA, & Hughes D (2015). The relationship between maternal education and children’s academic outcomes: A theoretical framework. Journal of Marriage and Family, 77(1), 60–76. [Google Scholar]

[R25] Henretta JC (2007). Early childbearing, marital status, and women’s health and mortality after age 50. Journal of Health and Social Behavior, 48(3), 254–266. 10.1177/002214650704800304 [DOI] [PubMed] [Google Scholar]

[R26] Hu L, & Bentler PM (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6(1), 1–55. 10.1080/10705519909540118 [DOI] [Google Scholar]

[R27] Huebner AJ, & Howell LW (2003). Examining the relationship between adolescent sexual risk-taking and perceptions of monitoring, communication, and parenting styles. Journal of Adolescent Health, 33(2), 71–78. 10.1016/S1054-139X(03)00141-1 [DOI] [PubMed] [Google Scholar]

[R28] Jaffee S, Caspi A, Moffitt TE, Belsky J, & Silva P (2001). Why are children born to teen mothers at risk for adverse outcomes in young adulthood? Results from a 20-year longitudinal study. Development and Psychopathology, 13(2), 377–397. 10.1017/S0954579401002103 [DOI] [PubMed] [Google Scholar]

[R29] Jutte DP, Roos NP, Brownell MD, Briggs G, MacWilliam L, & Roos LL (2010). The ripples of adolescent motherhood: Social, educational, and medical outcomes for children of teen and prior teen mothers. Academic Pediatrics, 10(5), 293–301. 10.1016/j.acap.2010.06.008 [DOI] [PubMed] [Google Scholar]

[R30] Kane JB, Philip Morgan S, Harris KM, & Guilkey DK (2013). The educational consequences of teen childbearing. Demography, 50(6), 2129–2150. 10.1007/s13524-013-0238-9 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R31] Keith TZ (2006). Multisample Models. In Multiple Regression and Beyond (1st ed., pp. 362–374). Boston, MA: Pearson Education, Inc. [Google Scholar]

[R32] Kim K (2014). Intergenerational transmission of age at first birth in the United States: Evidence from multiple surveys. Population Research and Policy Review, 33(5), 649–671. 10.1007/s11113-014-9328-7 [DOI] [Google Scholar]

[R33] Kirby D, & Lepore G (2007, November 26). Sexual risk and protective factors: Factors affecting teen sexual behavior, pregnancy, childbearing and sexually transmitted disease: Which are important? Which can you change? ETR Associates; Retrieved from http://recapp.etr.org/recapp/documents/theories/RiskProtectiveFactors200712.pdf [Google Scholar]

[R34] Kolenikov S, & Angeles G (2009). Socioeconomic status measurement with discrete proxy variables: Is principal component analysis a reliable answer? Review of Income and Wealth, 55(1), 128–165. 10.1111/j.1475-4991.2008.00309.x [DOI] [Google Scholar]

[R35] Lee D (2010). The early socioeconomic effects of teenage childbearing: A propensity score matching approach. Demographic Research, 23, 697–736. 10.4054/DemRes.2010.23.25 [DOI] [Google Scholar]

[R36] Lee Y (2009). Early motherhood and harsh parenting: The role of human, social, and cultural capital. Child Abuse & Neglect, 33(9), 625–637. 10.1016/j.chiabu.2009.02.007 [DOI] [PubMed] [Google Scholar]

[R37] Lewin A, Mitchell SJ, & Ronzio CR (2013). Developmental differences in parenting behavior: Comparing adolescent, emerging adult, and adult mothers. Merrill-Palmer Quarterly, 59(1), 23–49. 10.1353/mpq.2013.0003 [DOI] [Google Scholar]

[R38] Li X, Stanton B, & Feigelman S (2000). Impact of perceived parental monitoring on adolescent risk behavior over 4 years. Journal of Adolescent Health, 27(1), 49–56. 10.1016/S1054-139X(00)00092-6 [DOI] [PubMed] [Google Scholar]

[R39] Lindberg L, Santelli J, & Desai S (2016). Understanding the decline in adolescent fertility in the United States, 2007–2012. Journal of Adolescent Health, 59(5), 577–583. 10.1016/j.jadohealth.2016.06.024 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R40] Lou C, & Thomas A (2015). The relationship between academic achievement and nonmarital teenage childbearing: Evidence from the panel study of income dynamics. Perspectives on Sexual and Reproductive Health, 47(2), 91–98. 10.1363/47e2115 [DOI] [PubMed] [Google Scholar]

[R41] Luster T, & Small SA (1994). Factors associated with sexual risk-taking behaviors among adolescents. Journal of Marriage and the Family, 56(3), 622–632. 10.2307/352873 [DOI] [Google Scholar]

[R42] Magnuson K (2007). Maternal education and children’s academic achievement during middle childhood. Developmental Psychology, 43(6), 1497–1512. 10.1037/0012-1649.43.6.1497 [DOI] [PubMed] [Google Scholar]

[R43] Manlove J (1998). The influence of high school dropout and school disengagement on the risk of school-age pregnancy. Journal of Research on Adolescence, 8(2), 187–220. 10.1207/s15327795jra0802_2 [DOI] [PubMed] [Google Scholar]

[R44] Markwardt FC (1989). Peabody individual achievement test-revised Circle Pines, MN: American Guidance Service. [Google Scholar]

[R45] Martin JA, Hamilton BE, Osterman MJK, Driscoll AK, & Drake P (2018, January 31). Births: Final data for 2016 (National Vital Statistics Reports Vol. 67 No. 1). National Center for Health Statistics; Retrieved from https://www.cdc.gov/nchs/data/nvsr/nvsr67/nvsr67_01.pdf [PubMed] [Google Scholar]

[R46] Meade CS, Kershaw TS, & Ickovics JR (2008). The intergenerational cycle of teenage motherhood: An ecological approach. Health Psychology, 27(4), 419 10.1037/0278-6133.27.4.419 [DOI] [PubMed] [Google Scholar]

[R47] Miller BC (2002). Family influences on adolescent sexual and contraceptive behavior. Journal of Sex Research, 39(1), 22–26. 10.1080/00224490209552115 [DOI] [PubMed] [Google Scholar]

[R48] Muthén LK, & Muthén BO (2012). Mplus user’s guide seventh edition. Los Angeles, CA:Muthén & Muthén. [Google Scholar]

[R49] Perper K, Peterson K, & Manlove J (2010). Diploma attainment among teen mothers. Child Trends Retrieved from https://www.childtrends.org/wp-content/uploads/2010/01/child_trends-2010_01_22_FS_diplomaattainment.pdf

[R50] Pillow WS (2004). Unfit subjects: Educational policy and the teen mother New York, NY, US: RoutledgeFalmer. [Google Scholar]

[R51] Pogarsky G, Thornberry TP, & Lizotte AJ (2006). Developmental outcomes for children of young mothers. Journal of Marriage and Family, 68(2), 332–344. 10.1111/j.1741-3737.2006.00256.x [DOI] [Google Scholar]

[R52] Rai AA, Stanton B, Wu Y, Li X, Galbraith J, Cottrell L, … Burns J (2003). Relative influences of perceived parental monitoring and perceived peer involvement on adolescent risk behaviors: An analysis of six cross-sectional data sets. Journal of Adolescent Health, 33(2), 108–118. 10.1016/s1054-139x(03)00179-4 [DOI] [PubMed] [Google Scholar]

[R53] Romer D, Stanton B, Galbraith J, Feigelman S, Black MM, & Li X (1999). Parental influence on adolescent sexual behavior in high-poverty settings. Archives of Pediatrics & Adolescent Medicine, 153(10), 1055–1062. 10.1001/archpedi.153.10.1055 [DOI] [PubMed] [Google Scholar]

[R54] Ross CE, & Mirowsky J (2006). Sex differences in the effect of education on depression: resource multiplication or resource substitution? Social Science & Medicine, 63(5), 1400–1413. 10.1016/j.socscimed.2006.03.013 [DOI] [PubMed] [Google Scholar]

[R55] Ross CE, & Mirowsky J (2011). The interaction of personal and parental education on health. Social Science & Medicine, 72(4), 591–599. 10.1016/j.socscimed.2010.11.028 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R56] Sanchez D, Whittaker TA, & Hamilton E (2016). Perceived discrimination, peer influence and sexual behaviors in Mexican American preadolescents. Journal of Youth and Adolescence, 45(5), 928–944. 10.1007/s10964-016-0420-7 [DOI] [PubMed] [Google Scholar]

[R57] Schreiber JB, Nora A, Stage FK, Barlow EA, & King J (2006). Reporting structural equation modeling and confirmatory factor analysis results: A review. The Journal of Educational Research, 99(6), 323–338. 10.3200/joer.99.6.323-338 [DOI] [Google Scholar]

[R58] Strickland BB, Jones JR, Ghandour RM, Kogan MD, & Newacheck PW (2011). The medical home: Health care access and impact for children and youth in the United States. Pediatrics, 127(4), 604–611. 10.1542/peds.2009-3555 [DOI] [PubMed] [Google Scholar]

[R59] Sullivan CJ, Childs KK, & O’Connell D (2010). Adolescent risk behavior subgroups: An empirical assessment. Journal of Youth and Adolescence, 39(5), 541–562. 10.1007/s10964-009-9445-5 [DOI] [PubMed] [Google Scholar]

[R60] Tang S, Davis-Kean PE, Chen M, & Sexton HR (2014). Adolescent pregnancy’s intergenerational effects: Does an adolescent mother’s education have consequences for her children’s achievement? Journal of Research on Adolescence, 26(1), 180–193. 10.1111/jora.12182 [DOI] [Google Scholar]

[R61] US Bureau of Labor Statistics. (2014, June 5). Educational attainment and earnings of women Retrieved from http://www.bls.gov/opub/ted/2014/ted_20140603.htm

[R62] US Bureau of Labor Statistics. (n.d.-a). National Longitudinal Survey of Youth 1979 sample weights & clustering adjustments: Practical usage of weights Retrieved from https://www.nlsinfo.org/content/cohorts/nlsy79/using-and-understanding-the-data/sample-weights-clustering-adjustments

[R63] US Bureau of Labor Statistics. (n.d.-b). The NLSY79 Retrieved from https://www.bls.gov/nls/nlsy79.htm

[R64] US Bureau of Labor Statistics. (n.d.-c). The NLSY79 Child/Young Adult sample: An introduction Retrieved from https://www.nlsinfo.org/content/cohorts/nlsy79-children/intro-to-the-sample/nlsy79-childyoung-adult-sample-introduction

[R65] US Bureau of Labor Statistics. (n.d.-d). National Longitudinal Survey of Youth 1979 Children and Young Adults: Behavior Problem Index (BPI) Retrieved from https://www.nlsinfo.org/content/cohorts/nlsy79-children/topical-guide/assessments/behavior-problems-index-bpi

[R66] U.S. Census Bureau. (2015, June). Annual estimates of the resident population by sex, single year of age, race, and Hispanic origin for the United States: April 1, 2010 to July 1, 2014 Retrieved from http://factfinder.census.gov/bkmk/table/1.0/en/PEP/2014/PEPALL6N?slice=hisp~nhisp

[R67] Ventura SJ, Hamilton BE, & Matthews TJ (2014, August 20). National and state patterns of teen births in the United States, 1940–2013 (National Vital Statistics Reports Vol. 63 No. 4). National Center for Health Statistics; Retrieved from https://www.cdc.gov/nchs/data/nvsr/nvsr63/nvsr63_04.pdf [PubMed] [Google Scholar]

[R68] Wildsmith E, Manlove J, Jekielek S, Moore KA, & Mincieli L (2012). Teenage childbearing among youth born to teenage mothers. Youth & Society, 44(2), 258–283. 10.1177/0044118X11398366 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R69] Yu CY (2002). Evaluating cutoff criteria of model fit indices for latent variable models with binary and continuous outcomes Doctoral dissertation, University of California Los Angeles; Retrieved from https://pdfs.semanticscholar.org/7a22/ae22553f78582fc61c6cab4567d36998293b.pdf [Google Scholar]

[R70] Zajacova A, & Everett BG (2014). The nonequivalent health of high school equivalents. Social Science Quarterly, 95(1), 221–238. 10.1111/ssqu.12039 [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

Teen Mothers’ Educational Attainment and their Children’s Risk for Teenage Childbearing

C Emily Hendrick

Julie Maslowsky

Abstract

The Intergenerational Cycle of Teenage Childbearing

Teen Mothers’ Education as a Potential Pathway for Reducing the Intergenerational Cycle of Teenage Childbearing

Figure 1.

Figure 2.

Through Children’s Environmental Contexts

Through More Involved Parenting

Through Children’s Own Academic Aptitude and Performance

Teenage Childbearing among Children of Teen Mothers versus Children of Non-Teen Mothers

The Resource Substitution Theory

Teen Mothers Earn GEDs at Higher Rates

The Current Study

Methods

Data Sources

National Longitudinal Survey of Youth 1979 (NLSY79).

National Longitudinal Survey of Youth 1979 Children and Young Adults (NLSY79CYA).

The analytic sample.

Figure 3.

Measures

Table 1.

G2 teenage childbearing.

G1 educational attainment.

Mediators: G2 environmental and personal characteristics at age 14.

Household income.

Perceptions of parental monitoring.

Academic aptitude.

Mediators: G2 sexual risk behaviors in adolescence.

Age at first sex.

Unprotected sex.

Moderator: G1 Teenage Motherhood Status.

Control variables.

Analytic Approach

Model 1.

Table 2.

Model 2 & Model 3.

Model 4.

Model 5.

Model 6.

Results

Figure 4.

Direct Relationships Moderated by Teen Motherhood Status

G1 high school diploma and G2 household income at age 14.

G1 high school diploma and G2 age at first sex.

Direct Relationships Not Moderated by Teen Motherhood Status

G1 GED and G2 household income at age 14.

G1 degree types and G2 perception of parental monitoring.

G1 degree types and G2 academic aptitude.

G2 environmental and personal characteristics at age 14, sexual risk behaviors, and teenage childbearing.

Indirect Relationships Between G1 Degree Types and G2 Teenage Childbearing

Table 3.

Discussion

Mechanisms by Which Maternal Education Level Predicts Risk for Teenage Childbearing

Most mechanisms were similar for teen and non-teen mothers.

Few mechanisms were moderated by teen motherhood status.

Household income directly predicted risk for teenage childbearing.

The GED and the High School Diploma for Teen and Non-teen Mothers

Limitations

Directions for Future Research

Conclusions

Acknowledgments

Contributor Information

References

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases